INDEX

Explanations

words related to cognitive abilities and skills

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Bride

-0.79

bart

-0.59

 Schwe

-0.59

 Destination

-0.58

gone

-0.58

 Hanson

-0.58

algia

-0.58

 Alonso

-0.57

 Hein

-0.56

 Simone

-0.56

POSITIVE LOGITS

to

0.91

Reviewer

0.86

ologies

0.86

ibility

0.80

bodied

0.78

itud

0.73

Ability

0.73

ibilities

0.72

assisted

0.72

ively

0.70

Activations Density 0.038%

words related to cognitive abilities and skills

No Comments

No Known Activations